Data Quality Assurance in Construction Environmental Product Declarations for Whole Building Life Cycle Assessment

The construction industry strongly impacts climate change due to its contribution to environmental emissions. The rise in digital environmental product declaration (EPD), a valuable source of data for whole building life cycle assessment (WBLCA), poses significant risks to the reliability of sustainability assessments due to data quality assurance issues. In addition, there has been a lack of focus on product-level uncertainty estimation using EPD for WBLCA and exploration of the challenges and their cumulative effect on data quality assurance. This research builds on the work of researchers who have developed methods to assess EPD quality. However, these methods are not sophisticated enough to report data quality assurance quantitatively and incorporate the result in uncertainty analysis. This research aims to develop a methodology that assesses the data quality assurance of construction materials EPD to improve the reliability of WBLCA through the use of trusted data. The research adopts a mixed approach, including three major phases. The first phase entails the definition and measures of variables. The second phase includes assessing the significance of the challenges and data quality assurance (DQA) indicators, developing the DQA model, developing the intelligent knowledge-based decision support system (KB-DSS), utilising the DQA score for uncertainty analysis, and assessing EPD reporting. The third phase includes validating the research outputs through presentation and survey to experts. Content analysis was employed to analyse the qualitative data while several statistical techniques were employed to analyse the quantitative data collected from 169 respondents globally.

The research reveals the top-most challenge in LCA implementation for EPD development and EPD implementation for WBLCA, which includes ‘Problems with data availability and quality for LCA’ and ‘Poor quality of several underlying PCRs’, respectively. This suggests incorporating robust data quality mechanisms into the EPD development process and integrating proper guidance for developing product category rules (PCRs). Furthermore, verification was reported to be the most significant indicator of data quality assurance. The unique classification of the challenges and DQA indicators shows critical areas for stakeholders.

The relationship between the challenge in LCA implementation for EPD development, EPD implementation for WBLCA, and DQA indicators, as revealed by the PLS-SEM analysis, indicates that both categories of challenges strongly influence each other. In contrast, the ‘Challenges → Data Quality Assurance Indicator model’ reveals 39% certainty in improving DQA when the challenges are resolved. The ‘knowledge and policy awareness’ category of the challenges is critical, indicating the need for stakeholders to invest in capacity building regarding EPD development and WBLCA. Subsequently, a multi-level framework with tailored approaches to resolving the challenge categories was developed. The intelligent KB-DSS, including a DQA model based on the Fuzzy Synthetic Evaluation (FSE) method, was developed to support EPD data quality assurance assessment. The DQA model consists of eighteen (18) indicators in five groups. The ‘Data reliability’ group is the most significant indicator group. The data quality assurance index (DQAI) from the DQA model formed the basis of the DQA score computation. Following the development of the DQA model, the intelligent KB-DSS with embedded DQA model and action plans from literature was developed in Microsoft Visual Studio. In order to demonstrate the applicability of the intelligent KB-DSS and DQA score, a case study validation using the concrete element in the project was conducted, and the findings provided empirical evidence as regards the tool implementation for WBLCA reliability decision-making. Finally, all the research outputs have contributed significantly to theory, methods and practice. The study provided valuable insights into the challenges associated with EPD development and utilisation which is lacking in existing studies. Stakeholders can leverage this information to resolve the challenges in EPD development and utilisation. Furthermore, this is the first study to introduce the concept of data quality assurance in EPD context and develop a metric system to measure data quality including a decision support system that can help facilitate data quality assurance assessment process. These outputs are expected to enhance the quality of EPDs for WBLCA, support sustainable decision-making, and provide a foundation for future research on EPD data quality in sustainability assessments.